International Edition
Latest News
Health

California Startup Says It Administered First AI-Designed Virus to Human

Radical Numerics administers Omnii-1 virus to human patient As a potential scalable solution to counter antibiotic-resistant superbugs, a custom AI-designed virus has been introduced into a human patient for the first time by an artificial intelligence research lab…

California Startup Says It Administered First AI-Designed Virus to Human

Radical Numerics administers Omnii-1 virus to human patient

As a potential scalable solution to counter antibiotic-resistant superbugs, a custom AI-designed virus has been introduced into a human patient for the first time by an artificial intelligence research lab located in California has. Radical Numerics announced the treatment on Tuesday, though the intervention remains pending scientific scrutiny through a peer-reviewed study.

Should it be confirmed, the therapy called Omnii-1 could represent the initial instance of artificial intelligence-engineered bacteriophages being given to a person. Bacteriophages are viruses that specifically infect bacteria, offering a targeted alternative to traditional antibiotics. Antibiotic-resistant superbugs are projected to cause nearly 40 million deaths by 2050.

How Radical Numerics engineered the Omnii-1 phage therapy

Traditional phage treatment requires researchers to conduct a time-consuming search through vast environmental collections to find a natural virus that matches a patient’s specific bacterial infection. Clinicians extract a bacterial sample to find or design effective phages, a challenge that relies on understanding the genetic features required to kill particular bacterial strains. Phages execute this by binding to specific receptor molecules on the surface of target bacteria, entering the cell, and hijacking it to replicate before breaking it open.

California Startup Says It Administered First AI-Designed Virus to Human
Photo: independent.co.uk

Radical Numerics claims its AI-driven platform bypasses this hurdle by enabling rapid, on-demand generation of phages tailored to patient requirements. The startup states that its AI model is trained on diverse genomic sequences, phage genomes, and data measuring past phage-host interactions. This training allows the artificial intelligence to identify and design candidate phages for specific therapeutic needs. The company synthesizes these AI-designed variants and measures their viability and activity across bacterial strains in the laboratory.

Radical Numerics treats patient with Omnii-1 designs

The Omnii-1 designs were synthesized and administered to a patient suffering from an antibiotic-resistant infection who had very few remaining medical options. Radical Numerics stated that the treatment occurred under an authorized, physician-led compassionate-use pathway. The startup noted that the cocktail containing Omnii-1-designed phage variants eliminated the patient’s antibiotic-resistant isolate in vitro during laboratory testing.

A Startup Says AI-Designed Viruses Reached a Patient | The Daily AI Stew, 21 Oct 2026

Detailed clinical outcomes from the human administration have not been publicly disclosed, with the company citing patient privacy. Radical Numerics has noted that observations from a single patient don’t establish that its AI phage therapy is generally safe or effective, and the specific type of bacterial infection treated has not been publicly identified.

Frequently Asked Questions about the AI phage treatment

What are bacteriophages and how do they protect gut health?

Bacteriophages are viruses that target and infect specific bacteria. Unlike broad-spectrum antibiotics that tend to affect beneficial gut microbes, phages possess specificity that spares helpful intestinal flora.

California Startup Says It Administered First AI-Designed Virus to Human
Photo: theregister.com

How does the Omnii-1 model predict successful phage activity?

The AI model is trained on diverse genomic sequences, phage genomes, and historical phage-host interaction data. This computational training enables the algorithm to identify and design candidate phages for specific therapeutic requirements.

Why are specific clinical outcomes unavailable for this patient?

Radical Numerics withheld specific clinical outcome details and the exact nature of the patient’s infection, citing patient privacy.

About the author: Dr Natalie Singh - Health Editor

Board‑certified internal‑medicine physician and MPH. Natalie authored peer‑reviewed studies on infectious disease and served as medical editor. “Dr. Natalie Singh delivers evidence‑based health news, medical breakthroughs, and expert wellness guidance.”